Re: [OMPI users] Best way to send on mpi c, architecture dependent data type

2019-03-15 Thread Sergio None
Jeff, Yes, i know that C99 have these fixed standard library. The point is that many common used function, rand for example, return non fixed types. And then you need doing castings, typically to more big types. It can be a bit annoying. That's why I was looking for an OpenMPI way to do this

Re: [OMPI users] Best way to send on mpi c, architecture dependent data type

2019-03-15 Thread Jeff Squyres (jsquyres) via users
On Mar 15, 2019, at 2:02 PM, Sergio None wrote: > > Yes, i know that C99 have these fixed standard library. > > The point is that many common used function, rand for example, return non > fixed types. And then you need doing castings, typically to more big types. > It can be a bit annoying. >

Re: [OMPI users] Best way to send on mpi c, architecture dependent data type

2019-03-15 Thread Sergio None
Gilles, I'll keep it in mind, install 64bits ARM Rapsbian version next time I reinstall OS. But for now, I have too many things installed and configured. I'll try second option, I will build OpenMPI 4.0.1 with --enable-heterogeneous. Thank you very much for your help. Regards, Sergio. On 3/1

Re: [OMPI users] Best way to send on mpi c, architecture dependent data type

2019-03-15 Thread Sergio None
George, Yes, this is an option, create my own type library, but i need do a cover of MPI_Send, MPI_Receive, ... and other things . So was searching for an OpenMPI way, instead of doing it myself. I'll try the option that mention Gilles Gouaillardet: i will build OpenMPI 4.0.1 with --enable-het

Re: [OMPI users] Best way to send on mpi c, architecture dependent data type

2019-03-15 Thread George Reeke
On Fri, 2019-03-15 at 18:02 +, Sergio None wrote: > Jeff, > > Yes, i know that C99 have these fixed standard library. > > The point is that many common used function, rand for example, return > non fixed types. And then you need doing castings, typically to more > big types. It can be a bit a

Re: [OMPI users] Double free in collectives

2019-03-15 Thread Jeff Squyres (jsquyres) via users
On Mar 14, 2019, at 8:46 PM, Gilles Gouaillardet wrote: > > The first location is indeed in ompi_coll_libnbc_iallreduce() What on earth was I looking at this morning? Scheesh. Thanks for setting me straight...! -- Jeff Squyres jsquy...@cisco.com